Table 3.

Performance metrics of the classification procedures in the UKN1 test, the UKK2 test, and the combination ‘mean’.

			1-Fold Cmax				20-Fold Cmax a
Test	Data	Procedure	AUC	Accuracy	Sensitivity	Specificity	AUC	Accuracy	Sensitivity	Specificity
UKN1	Cytotoxicity		0.63	0.56	0.26	1	0.73	0.67	0.46	1
	Gene expression	SPS	0.78	0.64	0.43	0.94	0.82	0.62	0.38	1
		Top-1000	0.84	0.69	0.70	0.69	0.90	0.59	0.46	0.8
	Cytotoxicity and gene expression	SPS	0.84	0.79	0.70	0.94	0.90	0.90	0.83	1
		Top-1000	0.88	0.85	0.96	0.69	0.95	0.87	0.92	0.8
		RT-qPCR (SPS-like)	Not calculated	0.74	0.57	1	Not calculated	0.90	0.83	1
		RT-qPCR (Top-1000-like)	0.76	0.77	0.74	0.81	0.91	0.87	0.88	0.87
UKK2	Cytotoxicity		0.61	0.54	0.22	1	0.63	0.54	0.25	1
	Gene expression	SPS	0.87	0.77	0.61	1	0.83	0.69	0.58	0.87
		Top-1000	0.93	0.79	0.74	0.88	0.92	0.77	0.71	0.87
	Cytotoxicity and gene expression	SPS	0.9	0.9	0.83	1	0.87	0.85	0.83	0.87
		Top-1000	0.95	0.92	0.96	0.88	0.94	0.92	0.96	0.87
Combination ‘mean’	Cytotoxicity		0.63	0.56	0.26	1	0.73	0.67	0.46	1
	Gene expression	SPS	0.89	0.77	0.61	1	0.84	0.64	0.42	1
		Top-1000	0.94	0.79	0.74	0.88	0.96	0.79	0.75	0.87
	Cytotoxicity and gene expression	SPS	0.92	0.92	0.87	1	0.91	0.92	0.88	1
		Top-1000	0.95	0.92	0.96	0.88	0.98	0.95	1	0.87

Cytotoxicity = Only cytotoxicity data were considered for the calculation of the metrics, i.e., cytotoxic conditions were considered as positive and non-cytotoxic conditions as negative test results. Gene expression = only gene expression data were considered for the calculation of the metrics. Cytotoxicity and gene expression = all data for cytotoxicity, as well as for gene expression, were considered for the calculation of the metrics. AUC (area-under-curve) = for each possible cut-off used as threshold, predictions were made for each of the conditions based on which sensitivity and specificity were calculated. The ROC-curve (receiver operator characteristic) was obtained by plotting all pairs of (1-specificity) and sensitivity against each other. The AUC was determined as the area under this ROC-curve. Accuracy = ratio of correct predictions ((true negatives and positives)/(true and false negatives and positives)) (Supplementary Table S1). Sensitivity = ratio of detected teratogens (true positives/(false negatives + true positives)) (Supplementary Table S1). Specificity = ratio of detected non-teratogens (true negatives/(true negatives + false positives)) (Supplementary Table S1. ^a Including 10-fold C_max carbamazepine, 1.67-fold C_max VPA, and 1-fold C_max samples of leflunomide, phenytoin, teriflunomide, and vismodegib. RT-qPCR metrics used 1-fold C_max results instead of 20-fold C_max results for levothyroxine. Bold = best metrices for each test for the SPS- and top-1000-procedure and RT-qPCR.